Full Waveform Inversion Method For High-speed Rail Seismic Wave

On the basis of introducing and absorbing the advanced technology,Chinese high-speed rail has continuously achieved innovative breakthroughs and made remarkable accomplishment.In the near future,with the gradual improvement of the modern high-speed rail network,the high-speed rail network in china will enter a new era.Subgrade safety is the basic guarantee for the safe operation of high-speed trains.And the scholars have shown that the moving high-speed train can actually be considered as a new and repeatable "artificial seismic source" with broadband and discrete spectrum,and providing a significant basis for achieving the high-precision detection of the high-speed rail subgrade.In addition,the analysis of a great deal of high-speed rail seismic data shows that the high-speed rail seismic wave contains a wealth of geological environment information.In consequence,achieving the high-precision detection of shallow surface geological structures nearby the high-speed rail network by taking advantage of the high-speed train seismic source and fully excavating the effective information of the high-speed train seismic wave possess profoundly theoretical and practical significance.The full waveform inversion(FWI)constructs data fitting based on the observed wave field data and the wave field data modeled by the wave equation.Its core is to use the kinematics and dynamics information of the pre-stack seismic wave field,such as amplitude,travel time,phase,etc.,to construct the residual between the modeled wave field and the observed wave field,and realize the high-precision modeling of the physical parameters of the underground media by achieving the optimal match between modeled data and observed data based on optimization algorithms.After years of development,full waveform inversion has gradually become an important method for obtaining information on physical parameters of underground media,and is widely applied in seismic exploration.The high-speed train seismic source is different from the traditional geophysical seismic source,and its seismic wave field is affected by the excitation mechanism of the seismic source and the interaction of the underground media.In order to fully exploit the wealth of geological structure information contained in the high-speed rail seismic wave,effectively utilize the new and efficient high-speed train seismic source,and realize the detection of shallow surface geological structures nearby the high-speed rail network,we initially investigate the elastic FWI method for high-speed rail seismic wave,the paper including the following parts:1.The objective of full waveform inversion is to achieve the optimal matching of modelling data and observed data.Therefore,it is also important to generate the high-precision modelling data.For the numerical modelling of high-speed rail seismic wave,numerical errors are inevitably generated due to the use of difference operators to approximate differential operators,which is called numerical dispersion.In order to achieve the high-precision numerical modelling for high-speed rail seismic wave and reduce the impact of numerical dispersion on the results of numerical modelling and inversion,this paper investigate a optimized staggered-grid finite-difference method(SGFDM)based on the least-squares combination of square window function(LSCSW)to reduce the calculating error caused by difference operator approximating differential operator,and suppress numerical dispersion and improve the accuracy and efficiency of numerical modelling.2.When dealing with the complex seismic wave field generated by the complex high-speed train seismic source,it is necessary to investigate the excitation mechanism of high-speed rail seismic wave and the response characteristics and laws of high-speed rail seismic wave field to lay the foundation for achieving elastic FWI for high-speed rail seismic wave.Referring to the parameters of "Harmony" CRH5 high-speed train,we build a high-speed train load model in the viaduct system,and investigate the excitation mechanism of high-speed rail seismic wave to derive the high-speed train seismic source time function which can be applied to FWI.And we applied the optimized SGFDM based on LSCSW in performing numerical modelling for high-speed rail seismic wave,and analyze the response characteristics and laws of the high-speed rail seismic wave field.3.Relying on the high-speed train seismic source time function,on the basis of realizing the numerical modelling of high-speed rail seismic wave and the analysis of response characteristics and laws of the high-speed rail seismic wave field,considering the increase of the number of carriage of the high-speed train and the increase speed of the high-speed train,the high-speed rail seismic wave will appear crosstalk phenomenon,we take a single-carriage passing through the viaduct as an example,the elastic FWI for high-speed rail seismic wave is realized by using conjugate gradient method,which has great convergence speed and less computation and storage,and the elastic FWI results and analysis for various models are given.In this paper,the elastic FWI for high-speed train seismic source has been preliminarily explored,and obtain great results both for the anomalous body inversion of simple model and for the inversion of complex model.The results of elastic FWI are basically convergent,which shows that the elastic FWI can be used to accurately describe the underground structure and physical properties.The investigation shows that the elastic FWI for high-speed rail seismic wave is feasible,which can effectively extract the physical parameter information of the underground media from the high-speed rail seismic wave.In order to excavate the wealth of geological structure information contained in the high-speed rail seismic wave,and deepen the realization of underground space detection and engineering protection applications nearby the high-speed rail network,in the later stage,we need to improve the investigation of the elastic FWI for high-speed rail seismic wave.